Dear Lucas,
IMO = In My Opinion, IMEX = In My Experience
Interesting post. I think I agree and disagree with some of your comments. It is possible that Portugal is less flexible in its auditorial demands than elsewhere perhaps.
Without reading all this thread in detail, I suspect most of these posts are correct in different ways.
IMEX, Calibration has certain uniformly accepted basic requirements, ie choice of a reference standard, choice of a reference method, use of primary / secondary standards etc. However, within these items, there may be substantial interpretational differences depending on yr choice of reference. I will give two simple practical examples but with non-simple theoretical backgrounds - (a) some "standard" refs say you can calibrate the 100degC point on a themocouple based thermometer by insertion into "boiling" distilled water. Others sources specify must be located in the steam, eg using a kettle. Personally I hv found that there may indeed be significant differences but often also depending on the actual instrument. (b) Similarly the typical method for the 0degC point is insertion into "melting" ice/water. One of my local calibration companies consistently got results significantly different to my own (difference ca 0.2 - 0.5degC) and claimed was due to their use of distilled water for the ice and "highly" crushed ice for the actual test. Frankly I never believed them but used their certificate just the same and set the zero point as to my own results. A later company got results more closely matching my own and said they suspected the previous deviation was due to an electronic temperature simulation "black' box being used which is indeed another commercial "standard" method. I shall never know the truth of course.
In contrast, the terminologies validation / verification have a multitude of interpretational differences at both the principle and operational level, again depending on the source reference. For example the current standard interpretation of HACCP validation in USA (NACMCF) sets validation as a subset of verification unlike (I think) ISO 22000. The former method clarifies its usage by (from memory) including text like "for the purposes of this document, validation is defined as ........" The result is that the basic HACCP flow layout is somewhat re-arranged. This is a well-known feature although frequently not mentioned in subsidiary "standard" presentations. Codex and related authors hv issued various nice documents on this subject, some of which even discuss these variations. The result is that “it’s up to you”.
My conclusion is (a) there are various acceptable ways to calibrate and even more acceptable ways to validate/verify although local conditions may restrict the options (??) (b) you sometimes need to be very careful when comparing data of any kind, including the above, to avoid the "apples and oranges" effect. My first questions when claimed discrepancies occur in results are – (1) detailed procedure reference and (2) method of calibration if instrumentaion is involved. It is amazing IMEX how often companies use shortcut methods without knowing the method limitations, the justification being that it is the "standard" company method. It's true that procedures like AOAC can be impossibly lengthy so shortened versions are inevitable but then knowledge is also a requirement.
I use similar to the Arya method for scales. Invest in primary stainless steel master weights (externally calibrated yearly with certificate [demanded by auditor]) and then create sub-master secondary brass weights (cheaper) for routine use after using the primary ones to “standardise” a chosen balance(s) (the choice depending on the absolute level and sensitivity required by respective daily use. It’s true a significant (but mainly one-off ) initial cost is involved plus also some deep thinking as to one’s exact usage however the cost of paying someone to on-site calibrate a large range of weight scales can be substantial unless they offer a package deal. I sometimes hv it done professionally first for (particularly) high weight units and then copy their method for ongoing work. However I agree it can get complicated at levels like 100kg or <=a few mg.
(I personally find it equally difficult to reliably achieve routine checking of accuracy for, say, 200 units of 5kg scales. In some respects, analog balances are easier than digital.)
Rgds / Charles.C
